This article reviews the evidence about screening for breast cancer. Results are now available from four randomized trials--one quasi-experimental study and three case control studies. Only the oldest, the Health Insurance Plan (HIP) Trial, shows any evidence of mortality reduction in women from 40 to 49 years of age. However, the evidence from the HIP trial is weak and somewhat contradictory. Preliminary data from a further trial in Sweden and the National Breast Screening Study in Canada confirm my conclusion that the scientific evidence is insufficient at present to recommend mammography screening for women 40-49 years of age. This conclusion reinforces the decision of the U.S. Preventive Services Task Force, but contradicts the guidelines of the American Cancer Society and the National Cancer Institute.